JPH09283858A - Compd. semiconductor optical device manufacturing method and apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a method and an apparatus for easily selectively forming a structure having semiconductor layers of different band structure in the same plane, such as two-wavelength semiconductor laser arrays. SOLUTION: A second semiconductor layer 103 having a smaller band gap than that of a first semiconductor is locally formed on a substrate 101. including a first semiconductor. The substrate 101 is heated by a heat source having a radiation spectrum including absorption spectra of the first and the second semiconductors while third semiconductor layers 105a, 105b are epitaxially grown on the substrate 101, using an organic metal. The third layers 105a, 105b are different in growing temp. and hence different in band structure.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for manufacturing a compound semiconductor device, particularly a semiconductor optical device represented by a semiconductor laser.

[0002]

2. Description of the Related Art For the fabrication of compound semiconductor devices,
Selective growth technology is extremely important. The following are selective growth techniques for compound optical devices that have been reported so far.

(A) Yamada et al. (NTT): Japanese Patent Publication No. 5-34
In the 3801 metalorganic molecular beam epitaxial method, the temperature distribution is controlled by scanning the Ar laser beam on the substrate to form a temperature distribution, and the growth rate and composition of the film grown on the substrate are controlled according to the temperature distribution. Multi-wavelength array etc. are integrated.

(B) Sasaki et al. (Research and development of optical technology): JP-A-6-236849 In the gas source MBE (Molecular Beam Epitaxy) method, a temperature distribution is created by scanning an Ar laser beam on a substrate to re-create the surface. By changing the structure, semiconductors having different orientations are selectively grown on the substrate.

(C) Sasaki et al. (NEC): JP-A-6-2
32492 MOCVD (Metal Organic Chemical Vapor Depositio)
In the method n), semiconductor layers having different compositions are stacked on the substrate by using the dependence of the growth rate on the selective mask width.

[0006]

However, these have the following drawbacks. Regarding (a) and (b) (1) The throughput is poor and there is a problem in device accuracy. (2) It is difficult to monitor the surface temperature of the part where the laser is applied, and as a result, the controllability deteriorates.

Regarding (c) (1) Since a dielectric film or the like is used as a selective growth mask, it is inevitable to mix the mask itself or impurities adhering to the mask, so that high crystal quality cannot be obtained. (2) Since the selective mask after growth is damaged, the removing process is complicated and the cost is increased. (3) In the case of a semiconductor laser, it is not only necessary to form an aperture pattern with a narrow optical waveguide width (several μm), but also it is easy to induce lattice defects due to the stress of the selective mask.

Therefore, (1) a first object of the present invention is to selectively and easily form semiconductor layers having different band structures (specifically, a two-wavelength semiconductor laser array etc.) in the same plane, To provide a device (see Example 1).

(2) A second object of the present invention is to provide a method and a device which are capable of selectively and easily forming semiconductor layers having different band structures in the same plane and which are more reliable. (See Example 2).

(3) A third object of the present invention is to provide a method and apparatus for selectively and easily forming semiconductor layers having different band structures (specifically, a multi-wavelength semiconductor laser array etc.) in the same plane. To provide (see Example 3).

(4) A fourth object of the present invention is to easily manufacture a semiconductor laser having active layers having different band structures in the cavity direction, specifically, a polarization switching laser whose gain polarization characteristic is controlled. A method and apparatus for doing so are provided (see Example 4).

(5) A fifth object of the present invention is to facilitate a semiconductor laser having active layers having different band structures in the cavity direction, specifically, an optical device in which a semiconductor laser and a light intensity modulator are integrated. It is to provide a manufacturing method and apparatus (see Example 5).

[0013]

[Means for Solving the Problems]

(1) The invention for achieving the first object is to provide a means for locally forming a second semiconductor layer having a band gap smaller than that of the semiconductor on a substrate made of the first semiconductor, and the semiconductor substrate. Epitaxially growing a third semiconductor using an organic metal while heating it with a heat source having a radiation spectrum including absorption spectra of the first and second semiconductors (for example, directly heating the substrate with a heat source such as a heater). And an apparatus for manufacturing a compound semiconductor device. More specifically, the first compound semiconductor is InP, the second compound semiconductor is InGaAs or InGaAsP composed of a single layer, and the third compound semiconductor is InGaAs or InGaAs.
It is characterized by having a multi-layer structure made of P or InP. Further, the epitaxial growth means is characterized by using triethylgallium (TEG) as a part of the raw material.

(2) The invention for achieving the second object comprises means for locally forming a second semiconductor layer having a band gap smaller than that of the semiconductor on a substrate made of the first semiconductor, and The semiconductor substrate is heated with a heat source having a radiation spectrum including the absorption spectra of the first and second semiconductors (for example, the substrate is directly heated with a heat source such as a heater), and the third semiconductor is treated with an organic metal. A method and an apparatus for manufacturing a compound semiconductor device, which comprises means for utilizing epitaxial growth. More specifically, the first compound semiconductor is InP, the second compound semiconductor has a quantum well structure using InP, InGaAs, or InGaAsP, and the third compound semiconductor is InGaAs, InGaAsP. Alternatively, it is characterized by having a multilayer structure made of InP. Further, the epitaxial growth means is characterized by using triethylgallium (TEG) as a part of the raw material.

(3) Means for achieving the third, fourth and fifth objects of the present invention is to execute the above (1) or (2) and a combination of both.

The principle of the present invention will be described with reference to specific examples.
The growth rate and composition of InGaAs or InGaAsP are GaAs, InAs, and G, which are binary compounds.
aP, InP (InAs and GaAs for the former, In for the latter)
It is determined by the growth rate of As, InP, GaAs, and GaP). The growth rate of each is determined by a growth method using a gas source, for example, when a CBE method (Chemical Beam Epitaxy) is used, an organic metal (such as TMI (trimethylindium) or TEG (triethylgallium) which is a group III gas source ( It is determined by the decomposition efficiency of metal oxide) and the surface desorption rate of As and P which are group V elements. In the CBE method, this decomposition efficiency depends on the surface temperature of the substrate. T
EG has a higher degree of dependence than TMI, so TEG
By controlling the decomposition efficiency of InGaAs (or In
The composition and strain of GaAsP) can be controlled.

FIG. 2 shows the GaAs component and the In content with respect to the substrate surface temperature when InGaAs is epitaxially grown on the InP substrate by the CBE method using TMI and TEG as group III raw materials and arsine (AsH ₃ ) as group V raw material.
It is the one schematically showing the individual growth rate of the As component.
From FIG. 2, it can be seen that while the growth rate of InAs is constant with respect to the growth temperature, GaAs has a maximum value region (window region) with a width of about 40 degrees near 520 ° C. 480
The growth rate of GaAs decreases at temperatures lower than ℃
The decomposition efficiency of G decreases, and the growth rate of GaAs decreases at temperatures higher than 530 ° C. because As is desorbed from the surface. On the high temperature side, the window area can be enlarged by increasing the As molecular beam.

FIG. 3 shows the relationship between the substrate surface temperature and the degree of lattice mismatch by using InGaAs and InGaAsP (wavelength 1.1).
5 μm composition). InGaAs and InGaAsP are both In
The growth conditions are selected so as to be lattice-matched with P. As can be seen from FIG. 3, InGaAsP has a stronger surface temperature dependency and a smaller window width than InGaAs. From a different point of view, InGaAsP can control the degree of lattice mismatch with the surface temperature. Further, from the viewpoint of the energy gap, in InGaAs, the lattice mismatch is equivalent to the composition, so that the energy gap is equal on the low temperature side and the high temperature side if the lattice mismatch is the same. In contrast, InGaA
In sP, the shift amount to the low energy side is larger on the high temperature side than on the low temperature side even if the lattice mismatch is the same. This is because the cause of the low energy shift is due to lack of Ga on the low temperature side, whereas it is due to lack of As on the high temperature side, and the magnitude thereof is larger for As deficiency.

On the other hand, as a method of controlling the substrate temperature, as shown in FIG. 5 (b), a method of heating directly from the rear surface of the substrate 101 (without any intervening heater) (direct heating method) ). At this time, the temperature control is usually monitored by a thermocouple 203 or the like arranged on the back side of the substrate 101, and PID (proportional plus integral plus de
rivative) control. It is known that when CBE growth of InGaAs is performed on an InP substrate using this direct heating method, the surface temperature rises during the growth. This is because InP has a larger bandgap than InGaAs, so that the radiation component contributing to the temperature rise of InGaAs from the heater is transmitted and the surface temperature of InGaAs during growth is raised. Therefore, I
Even if the temperature of the back surface of the nP substrate is constant, the temperature of the InGaAs surface becomes higher than the temperature of the InP surface. FIG.
InGaAs growth film thickness and InG under certain growth conditions
It is an experimental result of the surface temperature relationship of aAs. from now on,
"The surface temperature rises linearly up to a thickness of 0.3 μm,
After that, it shows a saturation tendency. "

The present invention was devised from the above experimental facts. That is, a heat absorption layer (hereinafter also referred to as a distributed heat absorption layer) that is selectively formed on a substrate is formed, and a direct heating method (this is preferable, but from the above point, the first and second semiconductors (substrate And a heat absorption layer), the substrate may be heated by a heat source having a radiation spectrum including an absorption spectrum of InGaAs or InG by a CBE method or the like (a growth method of epitaxially growing an organic metal).
When a layer containing aAsP or the like is grown, the surface temperature rises in the distributed heat absorption region.
The composition and layer thickness of the layer containing GaAsP or the like can be different from that of the non-distributed heat absorption region.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1 (2LD Array) FIG. 1 is a schematic view for explaining the first embodiment. First, n
Type InP substrate 101 with a thickness of 0. 15 μm n-type InG
The aAs region 103 is selectively formed (FIG. 1B).
The forming method at this time is arbitrary, but in the present embodiment, S
Using the iO ₂ 102 as a selection mask (see FIG. 1A),
A part of the substrate 101 was etched, and selective growth by MOCVD was used there. In the case of the present embodiment, after removing the selection mask 102, the InP substrate 101 and InGaA are removed.
Although the thickness of the InGaAs layer 103 is controlled so that s and the like are in the same plane, they do not necessarily have to be in the same plane.
Instead of using selective growth, InGaAs or the like may be grown on the entire surface and then selectively etched. Alternatively, a combination of the above two methods may be used (that is, In
After GaAs or the like is grown on the entire surface, InGaAs or the like is selectively grown using SiO ₂ or the like as a selection mask).

Next, a laser structure is grown by using the CBE method. At this time, as a substrate heating method, as shown in FIG. 5, a heating mechanism such as a heater 202 and a substrate holder 201 (so-called indium-free holder) capable of directly heating the substrate 101 are used. In this embodiment, tantalum is used as the material of the heater 202, the temperature is controlled by the thermocouple 203 installed on the rear surface of the substrate 101, and the substrate surface temperature distribution is monitored by a pyrometer. At this time, the heater 202 includes the InP substrate 101 and InGaAs 103
It is a heat source having a radiation spectrum including the absorption spectrum of the semiconductor. Further, there may be a substrate holder which is not indium-free but has indium interposed between the holder and the substrate. In this case, it is not direct heating, but I
Any heat source may be used as long as it has a radiation spectrum including the absorption spectrum of the nP substrate 101 and the semiconductor such as InGaAs 103.

Before the laser structure growth, when the substrate 101 is set at 480 ° C. while applying a P (phosphorus) molecular beam, I
The temperature was 480 ° C. in the nP region 101, but InGaAs
In the area 103, the temperature rises due to the heat absorption effect, and
° C. This temperature distribution can be easily observed with a pyrometer or the like. In this state, the following laser structure is continuously grown.

An n-type InP layer (thickness: 0.5 μm) is grown as the first cladding layer 104, and I is formed as the active layer 105.
nGaAs well layer (thickness 4 nm) and InGaAsP
A quantum well layer (total well number N: 3) including a barrier layer (emission wavelength 1.15 μm composition, thickness 10 nm) and an InGaAsP carrier confinement (SCH) layer (emission wavelength 1.15 μm composition, thickness 50 nm) is grown. Further, as the second cladding layer 106, a p-type InP layer (thickness: 1.5 μm)
To grow p-type InGaA as the contact layer 107.
An s layer (0.5 μm thick) is grown. I of the active layer 105
The growth conditions for the nGaAs layer and the InGaAsP layer are 5
It was set so as to obtain lattice matching at 20 ° C. (corresponding to the conditions in FIG. 3).

As a result, first the first cladding layer 104 is
The distributed heat absorption region (InGaAs region 103) and the InP region 101 have the same layer thickness. This is TM
This is because there is no difference in the decomposition efficiency in the temperature range where I is 480 ° C. to 520 ° C. (see FIG. 2), and the thicknesses are also equal. Next, the active layer 105 has the same structure as the active layer 105b of the distributed heat absorption region 103 because the surface temperature is 520 ° C., but the active layer 105a of the InP region 101 has a significantly different structure. This is because the surface temperature of the InP region 101 is 480 ° C., so that the decomposition efficiency of TEG is low and the growth rate of the GaAs component forming InGaAs and InGaAsP is low (see FIG. 2). Therefore, as can be seen from FIG. 3, compressive strain is applied and the layer thickness is reduced. In the case of the present embodiment, as a result, the wavelength in the InP region 101 is 1.5 μm,
In the InGaAs region 103, active layers 105a and 105b having an oscillation wavelength of 1.3 μm were formed. In the present example, the thickness of the distributed heat absorption layer 103 was set to 0.15 μm, so the above result was obtained. It should be thicker.

Thereafter, in order to carry out lateral confinement,
After the mesa etching, the burying growth is performed, and the growth method may be selected according to the purpose. For example, instead of the above direct heating method, if CBE growth is performed by sticking to a molybdenum holder or the like with In, embedded growth without the influence of the distributed heat absorption layer 103 is possible. Indium free
The holder 201 may be left as it is, and after performing mesa etching, burying growth may be performed. After embedded growth, positive electrode 11
After the zero and negative electrodes 109 are formed, an element isolation groove 111 is formed in order to eliminate electrical and thermal crosstalk, thus completing the present embodiment.

The effects of this embodiment are as follows. (1) Active layers 105a and 10 of LD arrays having different wavelengths
5b can be produced by one crystal growth. (2) By controlling the thickness of the distributed heat absorption layer 103, the structure of the active layer 105 can be controlled, and the oscillation wavelength can be controlled.

Embodiment 2 (MQW distributed heat absorption layer) In the first embodiment, InGaA is used as the distributed heat absorption layer 103.
s monolayer was used. However, since InGaAs functions as a light absorption layer in an LD structure having a layer with a bandgap larger than that as an active layer, InGaAs may have a problem. In that case, this problem can be avoided by using a multiple quantum well layer as the distributed heat absorption layer 103 instead of bulk InGaAs. FIG. 6 shows an example of the multiple quantum well type temperature distribution layer 103. An n-type InGaAs layer (thickness 10 nm) is used as a well layer, and an n-type InP (thickness 10 is used) as a barrier layer.
nm) and the total number of wells may be selected according to the required heat absorption effect, but 10 was used in this example. The subsequent manufacturing process is the same as that of the first embodiment.

The effects peculiar to the second embodiment are as follows. (1) Since the MQW layer is used as the distributed heat absorption layer 103, light absorption here is reduced, so that the device design can be performed independently of the distributed heat absorption layer 103 (for example, the thickness of the first cladding layer 104). Can be small). (2) Since the distance from the distributed heat absorption layer 103 to the active layer 105 can be shortened, the heat absorption effect (on the active layer 105 etc.) of the distributed heat absorption layer 103 can be increased. (3) The distributed heat absorption layer 103 also has the effect of a superlattice buffer layer that prevents the growth of defects.

Example 3 (Multi-multi-wavelength LD array) The distribution of the surface temperature can be changed by changing the structure of the distributed heat absorption layer 103. FIG. 7 shows the distributed heat absorption layer 103.
The 3LD array in which the surface temperature distribution is controlled by changing the thicknesses of a and 103b and the structure of the active layers 105a, 105b, and 105c is changed is shown. First InGaAs region 1
03a is InGaAs with a thickness of 0.1 μm
In the nGaAs region 103b, In having a thickness of 0.15 μm is formed.
GaAs is embedded (FIG. 7A). By using the same growth conditions as in the first embodiment, active layers 105a, 105b, 1 having well layers having different strain amounts and different layer thicknesses are formed.
It is possible to obtain three kinds of 05c at the same time. Although the bulk InGaAs layer 103 is used here, the MQW having a different structure as in the second embodiment is used as the distributed heat absorption layer 103.
A layer (for example, a structure having different well layer thickness x number of wells) may be selected. Note that, in FIG. 7, the same reference numerals as those in FIG. 1 indicate the portions having the same functions.

Embodiment 4 (Polarization modulation with TE / TM region)
LD) In the above embodiments, an example of application to an arrayed LD is shown. Next, an example in which different active layer structures are formed in the resonator direction of a single LD will be shown. First, the polarization switching laser of this embodiment will be briefly described. FIG. 8 is a schematic view of the manufacturing process of the polarization switching laser.

As shown in FIG. 8D, the active layer 105 of this embodiment has a TE gain dominant region (TE region) 105a.
And a TM gain dominant region (TM region) 105b. By injecting carriers into these two regions independently, the TE mode and the TM mode of the Bragg wavelength compete with each other for polarization modulation, and high speed modulation or optical It is used for switches, etc.

In this laser manufacturing method, the distributed heat absorption layer 103 is formed in the cavity direction (not in the direction perpendicular to the cavity direction as in the above embodiment), and the distributed feedback reflector is formed as the cavity. Except for this, it is exactly the same as the above embodiment. In this embodiment, the MQW structure used in the second embodiment is adopted as the distributed heat absorption layer 103. Note that, in FIG. 8, the same reference numerals as those in FIG. 1 indicate the same functional portions.

The design guideline of this embodiment is as follows. (1) A non-strained quantum well is used in the active layer 105a in the TE region. (2) In the active layer 105b in the TM region, tensile strain is introduced into the barrier layer. (3) The gain peaks in the TE region and the TM region are made substantially equal.

Based on this design guideline, the growth conditions of the active layer 105 by the CBE method are as follows: TE on InP 101
A TM region is set on the region and the distributed heat absorption region 103 (FIG. 8A), and the number of wells of the distributed heat absorption layer 103 is adjusted so that when the surface temperature is 470 ° C. in the TE region, it becomes 500 ° C. in the TM region. (When the quantum well structure of the second embodiment is used, the number of wells N = 7). As shown in FIG. 8B, after the n-type InP 104 is grown to 1.0 μm, the multiple quantum well active layer 105 is grown. As the structure of the active layer, the structure of the first embodiment is used. However, in the active layer 105a in the TE region, both the well layer and the barrier layer are strain-free, and the active layer 10 in the TM region is
In 5b, the TEG and TMI flow rates were adjusted so that the well layer had no strain and the barrier layer had a strain (tensile strain) of −0.5%. The V / III ratio used was 2.0. FIG. 9 shows the relationship between the lattice mismatch and the substrate surface temperature corresponding to FIG. 3 at this time.

As a result, the TE gain is dominant in the InP region 101 and the TM gain is dominant in the InGaAs region 103. After this, the p-type InGaAsP optical guide layer 12
No. 1 was grown to 0.3 μm, the growth was temporarily stopped, and the grating 122 corresponding to the oscillation wavelength was formed on the surface. Then, the p-type InP clad layer 106 and the p-type InGaA were formed.
The s contact layer 107 is formed (FIG. 8C).

In this embodiment, the tensile strain is introduced only in the barrier layer of the active layer 105b in order to make the TM gain dominant, but it is also possible to introduce the strain only in the well layer or both the well layer and the barrier layer. Is.

The effects of this embodiment are as follows. (1) A region where the TE gain is dominant and a region where the TM gain is dominant can be simultaneously produced by one growth. (2) When it is necessary to control the light confinement coefficient of the TE mode and the TM mode, the structure of the distributed heat absorption layer 103 (bandgap etc.) and the layer thickness of the first cladding layer 104 are controlled to confine the light. The coefficient can be controlled.

Embodiment 5 (LD + Modulator) FIG. 10 shows an embodiment in which an intensity modulator is integrated in a semiconductor laser. This device guides laser light continuously oscillated in an LD section to an optical modulator and superimposes a signal on a reverse bias voltage to the optical modulator to extinguish the light.

The design guidelines of this embodiment are as follows. (1) In the LD region, compressive strain is introduced into the active layer 105a in order to achieve a low threshold and a high differential gain. (2) In the active layer (absorption layer) 105b in the modulator region, L
The band gap has a lower energy than that of the D oscillation light.
The basic structure is the same as that of the fourth embodiment except that the grating 122 is not formed in the optical modulator portion, but from the viewpoint of controlling the band gap rather than the polarization characteristics, the growth temperature for growing the LD structure is increased. As the above, by selecting the window region and growing the optical modulator at a temperature higher than the window region, a semiconductor layer having a large band gap difference can be obtained.

In the above embodiment, the distributed heat absorption layer is InGaAs or a multi-quantum well layer containing the same, but the present invention is not limited to this. If the semiconductor layer has a band gap smaller than that of the substrate, the same design guideline can be applied although the effect is large or small. Also,
By changing the set growth temperature for each layer and changing the temperature distribution, a device having a more complicated structure can be manufactured.

[0042]

The effects of the present invention are as follows. (1) Semiconductor lasers having different structures can be selectively manufactured without using a selective mask. 1-1) Since there is no selective mask, a high-purity crystal free from the influence of residual impurities can be obtained, so that the device has high reliability. 1-2) Since there is no selective mask, the mask peeling step after growth can be omitted, so that the cost is low.

(2) By changing the structure (bandgap, etc.) of the distributed heat absorption layer, not only the heat distribution but also the light distribution can be controlled. (3) A multi-wavelength LD array can be easily manufactured with high reliability. (4) A polarization switching LD can be manufactured easily and with high reliability. (5) An integrated optical device such as an LD plus intensity modulator can be easily manufactured with high reliability.

[Brief description of drawings]

FIG. 1 is a schematic diagram illustrating a manufacturing process of a first embodiment.

FIG. 2 is a diagram showing the relationship between substrate temperature and individual growth rate during InGaAs growth.

FIG. 3 is a diagram showing a relationship of a lattice mismatch degree with respect to a substrate surface temperature of InGaAsP and InGaAs.

FIG. 4 is a graph showing surface temperature rise and I during InGaAs growth.
The figure which shows the relationship of nGaAs film thickness.

FIG. 5 is a schematic diagram illustrating an example of a direct heating method.

FIG. 6 is a schematic diagram illustrating a second embodiment.

FIG. 7 is a schematic diagram illustrating a manufacturing process of a third embodiment.

FIG. 8 is a schematic diagram illustrating a manufacturing process of a fourth embodiment.

FIG. 9 is a schematic diagram illustrating the design guideline of the fourth embodiment.

FIG. 10 is a schematic diagram illustrating a fifth embodiment.

[Explanation of symbols]

 101 substrate 102 selective growth mask 103, 103a, 103b distributed heat absorption layer 104 first clad layer 105, 105a, 105b, 105c active layer 106 second clad layer 107 contact layer 108 buried layer 109 substrate side electrode 110 upper electrode 111 element isolation Groove 121 Optical guide layer 122 Grating 201 Substrate holder 202 Heater 203 Thermocouple

Claims (12)

[Claims] 1. A means for locally forming a second semiconductor layer having a band gap smaller than that of the first semiconductor on a substrate made of the first semiconductor, and absorption spectra of the first and second semiconductors. An apparatus for manufacturing a compound semiconductor optical device, comprising: a heat source having a radiation spectrum containing: and means for epitaxially growing a third semiconductor layer on the substrate using an organic metal while heating the substrate with the heat source. . 2. The compound semiconductor optical device manufacturing apparatus according to claim 1, wherein the heat source directly heats the substrate by a heat source such as a heater. 3. The first compound semiconductor is InP and the second compound semiconductor is InP.
InGaAs or I consisting of a single layer of compound semiconductor
nGaAsP and the third compound semiconductor layer is InGa
3. The compound semiconductor optical device manufacturing apparatus according to claim 1, which has a multi-layer structure made of As, InGaAsP, or InP. 4. The first compound semiconductor is InP and the second compound semiconductor is InP.
The compound semiconductor layer of InP, InGaAs or In
3. A compound semiconductor optical device according to claim 1, wherein the compound semiconductor optical device has a quantum well structure using any of GaAsP and the third compound semiconductor layer has a multi-layer structure composed of InGaAs, InGaAsP or InP. apparatus. 5. The apparatus for manufacturing a compound semiconductor optical device according to claim 1, wherein the means for epitaxially growing uses triethylgallium (TEG) as a part of the raw material. 6. The device is a 2LD array, multi-multiwavelength L.
6. The compound semiconductor optical device manufacturing apparatus according to claim 1, wherein any one of the D array, the polarization modulation LD, and the device in which the LD and the modulator are integrated is manufactured. 7. A second semiconductor layer having a band gap smaller than that of the first semiconductor is locally formed on a substrate made of the first semiconductor, and then the absorption spectra of the first and second semiconductors. A method for producing a compound semiconductor optical device, which comprises epitaxially growing a third semiconductor layer on a substrate using an organic metal while heating the substrate with a heat source having a radiation spectrum containing. 8. The method for manufacturing a compound semiconductor optical device according to claim 7, wherein the heat source directly heats the substrate by a heat source such as a heater. 9. The first compound semiconductor is InP and the second compound semiconductor is InP.
InGaAs or I consisting of a single layer of compound semiconductor
nGaAsP and the third compound semiconductor layer is InGa
9. The method for manufacturing a compound semiconductor optical device according to claim 7, which has a multi-layered structure made of As, InGaAsP or InP. 10. The first compound semiconductor is InP and the second compound semiconductor layer is InP, InGaAs or I.
has a quantum well structure using any of nGaAsP,
The third compound semiconductor layer is InGaAs, InGaAsP
Alternatively, the compound semiconductor optical device manufacturing method according to claim 7 or 8, wherein the compound semiconductor optical device has a multilayer structure made of InP. 11. The method of manufacturing a compound semiconductor optical device according to claim 7, wherein the step of epitaxially growing uses triethylgallium (TEG) as a part of a raw material. 12. The method according to claim 1, wherein any one of a 2LD array, a multi-multiwavelength LD array, a polarization modulation LD, and a device in which an LD and a modulator are integrated is manufactured.
A method for manufacturing a compound semiconductor optical device according to any one of 1.